A Review of the 1998 Overview Analysis of Randomized Adjuvant Tamoxifen Breast Cancer Trials

Peter M. Ravdin, M.D., Ph.D.

University of Texas Health Sciences Center, San Antonio


This section discusses the results from the 1995 Overview which were published in 1998. The results of the 2000 Overview are as yet unpublished, and have not been formalized. There were however, when presented in Oxford in September of 2000, no major changes in comparison with the 1995 analyses. 




The most recent Overview of the effects of tamoxifen in adjuvant trials has been published in the April 1998 issue of the Lancet(1). This is the third such overview published with earlier analyses being done in 1988(2) and 1992(3). This, the most recent Overview is perhaps the most valuable of the 3 for a number of reasons. 


1) This meta-analysis now includes several trials in which patients were randomized to get 5 years of tamoxifen. In the past, the analyses of the effects were dominated by the results of trials in which tamoxifen was given for 2 years or less. Since that time a large randomized trial of different tamoxifen durations shows that the appropriate duration is 5 years, and that shorter durations were inferior(4). Because the Overview review here also clearly shows that 5 years of tamoxifen is superior, most of this review emphasizes the 5 year results, and is more relevant to the clinical standards of today.  


2) In the past the proportional risk reductions were reported for recurrence, and also for combined recurrence and mortality.  It was not possible to cleanly see what the impact was on overall mortality. The present analysis is reported with 2 major endpoints; reduction in recurrence, and reduction in mortality.


3) There is now more than 10 years of follow-up on many trials, giving the analyses more statistical power. 


The overall conclusion of the most recent Overview analysis is that for women with estrogen receptor positive (or unknown and probably positive) early breast  cancer, adjuvant tamoxifen causes substantial reductions in risk or recurrence and death, irrespective of nodal status, age or menopausal status, and whether the patient received chemotherapy. 


The central questions addressed by the Overview analysis were:


How long should Tamoxifen be given?


 5 years appears superior to shorter durations.


Although the appropriate duration of tamoxifen can not be formally statistically addressed in trials in which there was not a head to head comparison of different tamoxifen durations, this can be inferred indirectly from the Overview results shown in Table 1. All durations of tamoxifen clearly benefited patients, but longer durations conferred more benefit. This is consistent with the Swedish trial which showed that 5 years of adjuvant tamoxifen was superior to 2 years(4). The results of the NSABP B-14 extension suggested that continuation of adjuvant tamoxifen beyond 5 years did not confer any additional benefit(5). Therefore, at this time the widely accepted standard of care for adjuvant tamoxifen duration is 5 years and this is supported by the current Overview.



Should Tamoxifen be given to estrogen receptor (ER) negative patients?


Although this question is open to further research, treating ER negative patients is not strongly supported by the current analysis. The question of whether estrogen receptor negative patients benefit from tamoxifen was addressed by the Overview. There was a clear benefit conferred by tamoxifen in ER positive patients, and no statistically significant benefit seen in the ER poor patients.  If there is a benefit conferred by tamoxifen in ER negative patients it is small. 



The Overview did examine whether tamoxifen seemed particularly beneficial in patients with very high levels of ER (ER++, > 100 femtomoles of ER per mg cytosol protein) as compared to patients who had 10-99 femtomoles of ER. 



The Overview also examined whether the use of progesterone receptor (PR) determinations allowed refinement of the estimates of benefit. This analysis was not ideal in that PR information was available on fewer of the patients, and because there were very few patients in some interesting subsets (ER negative, PR positive).  Thus, the analysis was restricted for patients who were ER positive and PR negative. These patients seemed to benefit to a similar degree to ER positive patients overall with reductions in the risk of recurrence of 46% (se 9) and mortality of 28% (se 11).



Does Tamoxifen benefit both premenopausal and postmenopausal patients?




Overall the effectiveness of tamoxifen did not seem to be strongly modulated by age. The one subset that did not seem to benefit as much was those patients 50-59 years of age.  No comment was made about this in the Overview analysis, and in the Overview no such effect was seen in adjuvant tamoxifen trials of 1 and 2 years duration. Also, patients 50-59 years of age did not seem to do statistically significantly different than those overall (who had a 47% and a 26% proportional risk reduction for recurrence or death). Thus the low degree of mortality reduction seen for 50-59 year old women in the trials with 5 years of tamoxifen might be interpreted as an artifact caused by the play of chance (some such deviations from the expected are expected when multiple subset analyses are done). 



Age is more reliably recorded than menopausal status, and thus analyses were done on the basis on age rather than menopausal status. The Overview reports that women 40-49 and 50-59 were further subdivided by menopausal status, and that subdivision did not affect age-specific results. 



Does Tamoxifen confer benefit for both Node Negative (NN) and Node Positive (NP) patients?




This question was addressed in the 1992 meta-analysis and the results of the current 1998 Overview were no different. The degree of benefit in terms of proportional risk reduction is the same in node negative and node positive patients.  



The absolute net benefit of tamoxifen would be expected to be less for patients with node negative disease because in patients with less of a risk of recurrence and death the proportional risk reduction afforded by adjuvant tamoxifen will be less. This effect was obvious for absolute mortality reduction of death. Clearly quite a substantial amount of net benefit can be expected by most node negative and node positive patients.




Do patients who get adjuvant chemotherapy still benefit from adjuvant tamoxifen?




The Overview analysis found that in adjuvant chemotherapy plus 5 years of tamoxifen was substantially better than chemotherapy alone. The 95 percent confidence intervals of the proportional benefits of tamoxifen with and without chemotherapy overlap, so it can not be stated that tamoxifen is actually more effective in women who received chemotherapy (and this was not the case for trials in which women received 1 or 2 years of adjuvant tamoxifen). Nonetheless it is clear that tamoxifen does not appear less effective in women who received adjuvant chemotherapy. 



Unfortunately the relatively low numbers of patients in the Tam + C versus C trials does not allow statistically reliable statements to be made about whether the additional benefit of tamoxifen was equally large in both pre- and postmenopausal women.  The analysis shown below suggests that there are not major differences, but because of large standard errors the interaction of age plus chemotherapy on the effectiveness is still uncertain. However, it does not appear greatly different. The results of soon to be reported trials such as Intergroup 0100 (comparing CMF with CAF, with and without tamoxifen) may help clarify this question. This result is also consistent with the results of the meta-analysis of adjuvant oophorectomy trials in which adjuvant oophorectomy clearly added to the effectiveness of adjuvant chemotherapy(6). 




Is Tamoxifen safe, does it perhaps even confer net health benefits?


Risk of endometrial cancer is increased.

Risk of second primary breast cancer is reduced ~40% over the next decade.

There is no effect on net non-breast cancer related mortality.


The Overview allows the magnitude of the potential risks and benefits of tamoxifen to be assessed. The data from the Overview suggests that the excess risk of endometrial cancer is approximately 1%, but that the excess risk of death due to endometrial cancer was more modest at ~0.2%.  This is consistent with previous reviews addressing this concern(7).


On the basis of a previously reported apparent excess of colorectal cancer in some adjuvant trials, the question of whether tamoxifen may increase colorectal cancer was addressed. In the Overview analysis of colorectal cancer cases there is not apparent excess in those patients receiving tamoxifen. The Overview analysis published in 1992 suggested that there might be a reduction of cardiovascular mortality due to tamoxifen. This reduction might be expected on the basis of favorable effects of tamoxifen on HDL/LDL ratios and total cholesterol(8). However this was not seen in this analysis, although modest effects on cardiovascular mortality were not excluded. All cause mortality (of which cardiovascular mortality was a major component) was not affected by tamoxifen. 


An ~40% reduction in contralateral breast cancer was again documented in patients getting 5 years of tamoxifen.




The Overview's strongest conclusion is that if the tumor has positive estrogen receptors, then adjuvant tamoxifen, for 5 years, should produce benefits, largely irrespective of age, menopausal status, nodal status, or prior adjuvant chemotherapy.


It was noted however that these were proportional benefits, and the absolute amount of benefit would depend on the baseline risk of the patient, and might be relatively small in patients with small localized tumors of good histological grade.


Overall the Overview strongly supported the use of adjuvant endocrine therapy, specifically adjuvant Tamoxifen. 


(1) Anonymous: Tamoxifen for Early Breast Cancer: an Overview of Randomised. Early Breast Cancer Trialists' Collaborative Group.  Lancet 351: 1451-1467, 1998.


(2) Anonymous: Effects of Adjuvant Tamoxifen and of Cytotoxic Therapy on Mortality in Early Breast Cancer:  An Overview of 61 Randomised Trials Among 28,896 Women. Early Breast Cancer Trialists' Collaborative Group.   N Engl J Med 319:1681-1692, 1988.


(3) Anonymous; Systemic Treatment of Early Breast Cancer by Hormonal, Cytotoxic, or Immune Therapy. Early Breast Cancer Trialists' Collaborative Group.  Lancet 339:1-15, 71-85, 1992.


(4) Swedish Breast Cancer Cooperative Group:  Randomized Trial of Two Versus Five Years of Adjuvant Tamoxifen for Postmenopausal Early Stage Breast Cancer.  J Natl Cancer Inst 88:1543-1549, 1996.


(5) Fisher B, Dignam J, Bryant J, deCillis A, Wickerham DL, Wolmark N, et al.: The Worth of Five Versus More than Five Years of Tamoxifen Therapy for Breast Cancer Patients with Negative Lymph Nodes and Estrogen Receptor-Positive Tumors. J Natl Cancer Inst 88:1529-1542, 1996.


(6) Anonymous: Ovarian Ablation in Early Breast Cancer: An Overview of Randomised Trials. Early Breast Cancer Trialists' Collaborative Group. Lancet 348: 1189-1196, 1996.


(7)Fisher B, Costantino JP, Redmond CK, Fisher ER, Wickerham DL, Cronin WM, Bowman D, Couture J, Dimitrov NV, Evans J, Farrar W, Kavanah M, Lickley HL, Margolese R, Paterson AHG, Robidoux A, Shibata H, Terz J.: Endometrial Cancer in Tamoxifen Treated Breast Cancer Patients:Findings from the National Surgical Adjuvant Breast and Bowel Project (NSABP) B-14.  J Natl Cancer Inst 86:527-537, 1994.


(8) Love RR, Wiebe DA, Newcomb PA, et al.:  Effects of Tamoxifen on Cardiovascular Risk Factors in Postmenopausal Women. Ann Intern Med 115:860-864, 1991.